The present invention relates to network management, and more specifically, to integrated device management over an Ethernet network.
In CLOS configurations, multiple network switches can be deployed into a full mesh network. In some configurations, switches are grouped together in switch boxes providing increased switching capabilities at a physical point. In a scaled out network, 32, 64, or 128 switch boxes, for example, may be connected together in a full mesh formation.
Apart from network switching ASICs, network switch boxes typically also include a processor to control peripheral hardware devices that supply switching related functions. The processor is typically connected to local peripheral hardware through local bus lines. The processor configuration may be mapped directly to the hardware devices (for example, to device addresses) in the switch and the different port configurations. For example, the processor may be configured to enable ports, manage IP interfaces, and manage protocols. A switch box may typically include one processor that operates as a management plane and control plane, providing management capabilities and controlling hardware, protocols. Thus, in scaled out systems, a management point may typically be confined to communicating with local devices and there may be several management points requiring maintenance and monitoring of the local devices at each switch.
In network management, troubleshooting for a switch box typically occurs at the local level. An interface may be connected to each management point through a port in the switch. An administrator or technician may use a scripting mechanism through the interface to configure the processor in each switch independently. Once configured, the management point may be interfaced to control and monitor the performance of components, for example, ports in the switch. The management point may be accessed to request, for example, logs, port management data and health status of devices connected to the switch.
For each switch box in the network that requires maintenance, the administrator or technician typically has to troubleshoot a switch at the switch's physical location. An administrator may thus be challenged to provide real time services in a scaled out network where switch boxes may be numerous and some may be in physically distant locales. For example, when an interrupt occurs, the processing environment may crash requiring immediate attention from an administrator. Otherwise applications being handled by the network may not continue and transactions related to those applications can be lost. Some events, for example, a cooling fan becoming inoperable, may require imminent action before overheating occurs, however, when the fan is on a remote switch, the administrator may need to go on site to remedy the situation. As described above, when setting up a scaled out network, each switch box may require independent set-up and configuration. The manpower and time needed to set up and manage a scaled out network may thus, be cost ineffective.